Was the Black Sea Catastrophically Flooded in the Early Holocene?
Total Page:16
File Type:pdf, Size:1020Kb
Rapid Communication Was the Black Sea catastrophically flooded in the early Holocene? Liviu Giosana, Florin Filipb, Stefan Constatinescub aCoastal Systems Group, Department of Geology and Geophysics, Woods Hole Oceanographic Institution, 360 Woods Hole Road, Woods Hole, MA 02543, USA; Corresponding author: [email protected]; +1 508 289 2257. bDepartment of Geography, University of Bucharest, N. Balcescu No.1, Bucharest, Romania Abstract A catastrophic flooding of the Black Sea basin was proposed to have occurred during its reconnection to the ocean in the early Holocene. Possible cultural consequences of the flood include the migration of Neolithic farmers from around the Black Sea towards central Europe as well as the creation of flood myths. Stratigraphic and paleo- geomorphologic information from Danube delta aided by radiocarbon ages on articulated mollusks constrain the level in the Black Sea before the marine reconnection to ca. 30 m below the present sea level rather than 80 m or lower. If the flood occurred at all, the sea level increase and the flooded area during the reconnection were significantly smaller than previously proposed. 1. The last Black Sea reconnection to the ocean From early ideas (Murray, 1900) to modern studies, after a century of research, there is still little consensus on the Black Sea level variations (e.g., Pirazzoli, 1991; Ryan, 2007; Hiscott et al., 2007). The Black Sea is a marginal basin that becomes marine only when the ocean level rises above the depth of the straits connecting it to the Mediterranean: the Bosphorus (current depth: ~35 m) and the Dardanelles (current depth: ~70 m). During the last global lowstand and for much of the subsequent deglacial ocean level rise, the Black Sea functioned as a giant lake with its level controlled by climate (e.g., Ross et al., 1970). Contrary to previous research envisioning a smooth reconnection the ocean (e.g., Ross et al., 1970), Ryan and colleagues proposed that the Black Sea was instead catastrophically flooded and its level rose at least 50 meters from around 90 m or lower below present sea level in a few years time (Ryan et al., 1997; Ryan and Pitman, 2000). The authors suggested that flooding of the vast continental shelf may have led to a migration of Neolithic farmers from around the Black Sea towards central Europe and led to the creation of flood myths. A protracted debate continues on the occurrence of the flood itself as well as on its possible cultural consequences (e.g., Aksu et al., 2002; Ryan et al., 2003; Hiscott et al., 2007; Ryan, 2007; Turney and Brown, 2007; Yanko-Hombach et al., 2007a,b). Faunal and geochemical reconstructions converge to indicate a rapid transition from a fresh to brackish lake to the ocean-connected modern Black Sea around 9 400 years ago (~8 400 14C years BP; Ryan et al., 2003; Major et al., 2006; Hiscott et al., 2007; Bahr et al., 2008), although a weak or transitory inflow of Mediterranean waters before that time is still discussed (e.g., Hiscott et al., 2007). Water level variations accompanying these changes remain, however, contentious. An agreement exists that a lake highstand at ~20- 30 m below current sea level (mbsl) was reached sometimes during the deglacial to the earliest Holocene, but the exact timing, cause, and temporal extent of this is disputed (e.g., Ryan et al., 2003; Hiscott et al., 2007; Lericolais et al., 2007a). 2. Constraints on the Black Sea level changes Proponents of the flood hypothesis advanced a series of arguments in support of a lake level below the Bosphorus threshold immediately preceding the marine reconnection. At several locations on the Black Sea shelf, a pre-Younger Dryas lowstand (<13 250 14C years BP; Major, 2002) was found to be expressed as an unconformity at the top of desiccated clay-rich deposits with plant remains (e.g., Major, 2002; Ryan et al., 2003). A sandy coquina layer composed of lacustrine mollusks overlies lowstand deposits from just below 100 mbsl and was interpreted as a Younger Dryas trangressive lag (Major, 2002; Ryan, 2003). Supported by bulk carbonate ages of Dimitrov (1982) from the mid- shelf, Major (2002) suggested that the Younger Dryas coquina extended on the shelf shallower than 100 mbsl and was subaerially reworked during a post-Younger Dryas regression. The resulting shell hash was colonized from ~ 90 mbsl to higher than 40 mbsl by more salt-tolerant mollusk species after the reconnection (Major et al., 2006). At depths between 100 mbsl and 70 mbsl, drowned paleo-shoreline features and fields of bedforms, interpreted based on their morphometry as subaerial dune fields (Ryan et al., 2003; Ryan, 2007; Lericolais et al., 2007a, b), were also used to argue for a post-Younger Dryas regression, but these features are yet to be directly dated. Supporters of the continuous transgression hypothesis argued that bedforms preserved on the Black Sea shelf in front of the Bosphorus represent barrier islands and associated lagoonal deposits that were drowned as the lake level rose over the outer and mid shelf (Aksu et al., 2002). Subsequent analysis showed that the bedforms were instead part of a “shelf fan” constructed by the dense, bottom-hugging inflow of Mediterranean waters spreading on the shelf from the Bosphorus outlet after the ocean reconnection (Ryan, 2007) that may overlap older erosional relief (Giosan et al., 2005). However, in a core retrieved from 69 mbsl on the southwestern shelf of the Black Sea, sediments interpreted to have accumulated below wave base (taken as ≥ 30 m water depth) span the postulated age of the flood covering the entire interval from before 9 100 to after 8 300 14C years BP (Hiscott et al., 2007). If we assume that the region around Sakarya River estuary in Turkey was vertically stable over the Holocene, the Black Sea level after the reconnection to the ocean was >18 mbsl after 7 500 BP, the youngest calibrated age for lacustrine sediments before estuarine sedimentation started at that location (Görür et al., 2001). Ancillary support for a lacustrine highstand above the Bosphorus threshold comes from deltaic deposition in the Sea of Marmara from a proposed Black Sea water outflow between ~10 000 and 9 000 14C years BP (e.g., Hiscott et al., 2002), but it is highly contested with conflicting ages and different morpho-stratigraphic interpretations (e.g., Gökaşan et al., 2005; Eriş et al., 2007, 2008; cf., Hiscott et al., 2008). Other reconstructions have argued for an elevated lake level in the Black Sea, close to the modern depth of the Bosphorus, at the time of its reconnection to the ocean (e.g, Chepalyga, 1984; Balabanov, 2007; Yanko-Hombach et al., 2007a); however, documentation on the indicative meaning of the stratigraphic or morphologic features used as sea level indicators, on the accuracy of their chronologies, as well as on possible tectonic influences is needed to establish the value of these reconstructions (Pirazzoli, 1991; Giosan, 2007). 3. A Danube delta perspective The uncertainty surrounding the Black Sea level stems from a lack of reconstructions based on reliable sea level markers (Pirazzoli, 1991; Giosan et al., 2006), a scarcity of radiocarbon ages on in situ materials and the difficulty in calibrating radiocarbon ages in a setting with variable reservoir ages (Giosan, 2007; Ryan, 2007; Kwiecien et al., 2008). Compared to other coastal settings around the Black Sea, the Danube delta was vertically stable since at least the last interglacial (Dodonov et al., 2000; Giosan et al., 2006). A sea level reconstruction in the Danube delta shows that the Black Sea level was close to its modern level for the past five millennia (Giosan et al., 2006). Whereas the evolution of the Danube delta since the middle Holocene is more or less known (Fig. 1), earlier deglacial to early Holocene deltaic-estuarine phases remain to be deciphered. Sedimentary deposits immediately underlying the modern delta down to the coarse gravelly fluvial base (Fig. 2a) had been postulated to be as old as early to middle Pleistocene (Liteanu et al., 1961; Liteanu and Pricajan, 1963) to early Holocene (Panin, 1972). Recently, Lericolais et al. (2007c) seismically imaged a clinoform above an erosional surface at 47 mbsl, in front of the present delta coast (Fig. 2; Minereau, 2006), and inferred it to have formed during a Bolling-Allerod lacustrine highstand. FIGS. 1 and 2 here Numerous Danube delta drill cores from the late 1950s (Fig. 2) were discussed in early papers (Liteanu et al., 1961; Grossu and Baltac, 1962; Liteanu and Pricajan, 1963; Baltac, 1963, 1964; Panin, 1972, 2007; Romanescu, 1996), but core material has not been preserved to allow further study. In 2007, we drilled a new core to 42 m depth near the present deltaic coast, onshore of the recently surveyed pre-modern clinoform. Facies analysis based on lithology, sedimentary structures and textures, and a high resolution geochemical grain size proxy record, were employed to identify depositional environments (Fig. 3). Radiocarbon ages on articulated mollusks (Table I; Fig. 3) collected from sections with no drilling disturbance (i.e., intact stratification, preserved bedding contacts) provide age control. The dated mollusks were estimated to be in situ based on the lack of abrasion and secondary encrustations, non-exotic character (i.e., the lithological and facies characteristics of sediments that preserved them is in agreement with their known modern habitat; Nevesskaya, 1965; Bacescu et al., 1971), and sediment filling similar to the sediment surrounding the shells. We calibrated the radiocarbon ages using Calib 5.0.1 (Reimer et al., 2004); for marine mollusks we assumed a modern value for the reservoir age is 440 ± 40 yr BP with ΔR 75 ± 60 yr (Siani et al., 2001), whereas older ages on lacustrine mollusks were calibrated assuming a null reservoir age for the lacustrine surface waters after the Younger Dryas (Ryan, 2007; Kwiecin et al., 2008).